Characterization of The Mechanics of Compliant Wing Designs For Flapping-Wing Miniature Air Vehicles
Characterization of The Mechanics of Compliant Wing Designs For Flapping-Wing Miniature Air Vehicles
Characterization of The Mechanics of Compliant Wing Designs For Flapping-Wing Miniature Air Vehicles
DOI 10.1007/s11340-013-9779-5
Received: 10 September 2012 / Accepted: 24 June 2013 / Published online: 9 July 2013
# Society for Experimental Mechanics 2013
Abstract Flapping-wing miniature air vehicles (MAVs) of- wing, rotary-wing, and flapping-wing designs [1–9]. The large
fer multiple performance benefits relative to fixed-wing and variety of applications for MAVs reveals both advantages and
rotary-wing MAVs. This investigation focused on the prob- disadvantages to each type of wing design. However, the
lem of designing compliant wings for a flapping-wing MAV versatile flapping-wing design, which is biologically inspired,
where only the spar configuration was varied to achieve offers potential in a wide range of missions overlapping with
improved performance. Because the computational tools the capabilities of both fixed-wing and rotary-wing designs.
needed for identifying the optimal spar configuration for The field of flapping-wing miniature air vehicles is rela-
highly compliant wing designs have yet to be developed, a tively new, and is constantly growing and expanding. There
new experimental methodology was developed to explore are many designs currently being explored, with each one
the effects of spar configuration on the wing performance. emphasizing a different aspect of the vehicle’s construction
This technique optically characterized the wing deforma- [10]. Because the size range of unmanned flapping-wing air
tions associated with a given spar configuration and used a vehicles is so large, “miniature” with respect to unmanned
customized test stand for measuring lift and thrust loads on aerial vehicles (UAVs) constitutes any flier weighing be-
the wings during flapping. This revealed that spar configu- tween 10 and 100 g. Flapping-wing MAVs can be divided
rations achieving large and stable deformed volume during into three major categories that are critical in determining an
the flapping cycle provided the best combination of lift and MAV’s flight envelope: directional control scheme, mecha-
thrust. The approach also included a sensitivity and repro- nism design, and wing design [10].
ducibility analysis on potential spar configurations. Results Flapping wings are often constructed using stiff, lightweight
indicated that the wing shape and corresponding lift and rods as structural materials and a thin polymer film as the wing
thrust performance were very sensitive to slight changes in surface [11–14]. The wings of two types of birds designed at
volume and 3-D shape associated with slight variations in the the University of Maryland, a ‘Small Bird’ weighing approx-
spar locations. imately 12 g and a ‘Big Bird’ weighing approximately 30 g, as
well as the ornithopter developed at the University of Dela-
Keywords Miniature air vehicles . Shape characterization . ware, are all constructed by hand using this technique. The spar
Compliant wings . Lift and thrust measurement . Spar and film construction style is advantageous because of the
location ability to adjust the wing properties easily. Adjustment to the
stiffness distribution across the wing surface is accomplished
by changing the size and location of the stiffening spars.
Introduction However, because these wings are man-made, repeatability in
manufacturing is a challenge, as small differences are present
Currently, miniature air vehicles (MAVs) are being developed between each set of wings. Commercial MAVs, such as the I-
with variations on three different wing configurations: fixed- fly Vamp and Wasp within the toy market, also employ thin
films stretched across front and rear spars in their flapping
wings.
J.W. Gerdes : K.C. Cellon : H.A. Bruck (*) : S.K. Gupta An understanding of current wing design is an important
Department of Mechanical Engineering, University of Maryland,
College Park, MD 20742, USA focus of MAV development because it influences how the
e-mail: bruck@umd.edu wings are modified to optimize performance and informs the
1562 Exp Mech (2013) 53:1561–1571
Rigid links
Rigid body
Crank Fig. 4 MAV mounted in wind tunnel during aerodynamic thrust testing
Revolute joints
at 0° angle of attack
t1
1
H
t2
2
3
4
t5 t3
t4
Fig. 3 Test stand used for aerodynamic and static load measurement Fig. 5 Wing configuration template [27]
1564 Exp Mech (2013) 53:1561–1571
Characterization of Mechanics of Compliant Wing The general approach to the layout of the spars is seen in
Designs Fig. 5, with each wing being constructed with the same
BoPET-based foil pattern. The semi-span L was set to
Design of Compliant Wings 33.0 cm and the maximum chord H was set to 15.2 cm for
every wing design, resulting in an overall surface area of
The wings of any flapping-wing MAV are one of the most 502 cm2 for each wing. Six prototypes were constructed;
crucial components governing overall performance in flight. their specific parameters can be seen in Table 1. As these
Wings affect the endurance, speed, maneuverability, climbing, wings were hand-built, slight deviations inevitably occurred
gliding, and other flight behaviors. The wings are also the during construction, but the actual effect of the deviations
source of lift and thrust, providing the capability to support the had not been previously explored.
weight of the MAV and any additional desired payloads.
Flexible flapping wings deform as they are accelerated Lift and Thrust Measurements
through the flapping cycle into an airfoil shape and aerody-
namic loading produces large angles of attack to generate lift A good combination of lift and thrust values are necessary to
and thrust when in a moving airstream [26, 27]. achieve successful flight. The best performing wings have
The wings used in the ‘Jumbo Bird’ MAV were constructed large lift and thrust values; in a lift vs. thrust plot, the most
in a biologically-inspired style. A stiff skeletal structure of successful wings fall in the upper right quadrant. Using the
carbon fiber rods provided support. Desired stiffness distribu- methods described in Experimental Procedures section, the
tions were achieved by changing the diameters and orientation six wing configurations were tested in both lift and thrust
of the spar structure along the wing surface. A thin film of modes in the wind tunnel. The average lift and thrust results
Biaxially-oriented Polyethylene Terephthalate (BoPET) obtained over a flapping cycle during wind tunnel testing are
provided the lifting surface. Because the film was coated shown in Fig. 6. Wing F performed the best, maximizing lift
with a thermally-activated adhesive, no glues or adhesives and thrust, while wings A, B, D and E performed the worst,
were necessary in the assembly of the spars and film, a small with low values of lift and/or thrust.
but important detail in ensuring the smooth deformation of the In order to determine the validity of the measurements made
wings under loading. using the wind tunnel, outdoor flight testing was conducted.
Results can be seen in Fig. 7. The actual flight payload was
consistently on the order of 15–25 g greater than the lift
measured by the test stand because the wind tunnel was unable
Characterization of Wing Deformations center line during the transition between the up-flap
and the down-flap, and then move away from the center
Front Spar Motion line towards the rear of the wing once again during the
down-flap as the wing fills up with air and deforms,
Throughout the wing cycle, there are a few components that snapping back to zero at the transition. From observing
affect how the wing develops, one of which is the motion of the plot of the front spar paths, it appeared that wing F
the front spar. Using high-speed imaging, the movement of best followed the model path which allowed the wing to
the front spar was tracked throughout the flapping cycle from collect the most useful amount of air and expel it in the
two orthogonal perspectives: (a) from the front (head-on) most efficient way. Wings A and D followed the model
and (b) above the MAV (over-top) (see Fig. 11). From these path the least accurately, which may have attributed to
two perspectives, front spar tip displacement was quantified their generation of smaller lift and thrust loads.
by comparison with a superimposed perfectly stiff wing. The The model path for the front spar from the head-on per-
measurement was made using image analysis software on spective resembles a parallelogram; the leading spar trails the
each of the fifteen evenly spaced frames for each of the six model spar location during the up-flap so the displacement is
different wing types. negative, crosses over the model location during the transi-
Plots comparing the front spar displacement for both the tion between up and down flaps, and then trails the model spar
over-top and head-on perspectives can be seen in Figs. 12 location during the down-flap, resulting in positive displace-
and 13, respectively. The path of the front spar from both ment. As seen in Fig. 13, all of the wing designs followed the
perspectives are consistent with models of lift and thrust model front spar path in approximately the same manner; this
profiles presented in previous work [25] are shown in the reflected the fact that the wings are all constructed with the
figures using dashed lines, as well as the lift results in Fig. 9 same leading spar thickness, 1.5 mm carbon fiber.
(correlates with head-on) and the thrust results in Fig. 10 The displacement of the leading spar, from both the head-
(correlates with over-top). From the over-top perspective, the on and over-top perspectives, indicated the direction in
front spar should move away from the center line towards the which the spar moved during the up-flap and down-flap of
rear of the wing during the up-flap, snap forward of the the flapping cycle. This improved the understanding of wing
shape during the cycle and how that shape affected the lift again as the wing moves through the down-flap and air is
and thrust loads generated during flight because it defined collected, to be expelled once again on the transition to the
the location of the front edge of the wing at fifteen different up-flap. Based on the results, it is obvious that wing F collected
points throughout the flapping cycle. significantly more volume than the other wing designs. This
reinforced the fact that wing F was the top-performing wing out
Volume Measurement of the six; its lift and thrust data reflected the large amount of
volume displaced by the wing design.
Volumes were measured at fifteen evenly spaced points
throughout the flapping cycle for all six wing configurations. Repeatability and Sensitivity Analysis
Figure 14 compares the normalized volume over the flapping
cycle for all six configurations. Ideally, the wing volume should A sensitivity analysis for the measured volume was also
increase on the up-flap as the aerodynamic loads cause the wing performed for the four identical wings at fifteen different
to billow towards the ground, come back down through the angles throughout the flapping cycle, as was done previously
transition between up and down-flap, and then quickly grow for the lift and thrust measurements. Figure 15 compares the
again as the down-flap is initiated. The volume should increase progression of volume captured throughout the flapping
cycle for the four wings. As seen in this figure, the volume calculated from the in-plane deformations in the x-direction
captured by the four wing sets was similar; however, the two and y-direction (directions relative to the wing can be seen in
wings in which the 1 mm spar was moved, named sensitivity Fig. 11), and then all three deformations were combined to
in and sensitivity out in the graph, captured less volume than obtain the 3-D shape. Therefore, all of the grid points in each
the other three wing sets. Also, the two wings identical to the wing image were tracked throughout 15 different mechanism
original F wing design captured slightly less volume than the angles. Using image analysis software, coordinates for each
original wing, but still captured more volume than the other point in the wing throughout the 15 different mechanism
two wing sets. Since the difference in volume was small, the angles could easily be determined. Figure 16 shows the mea-
differences in the lift and thrust values must come from sured 3-D shapes for wing A, and Fig. 17 shows the measured
another factor, such as the overall shape of the wing through- 3-D shapes for wing F.
out the flapping cycle. Much like the calculated images, the 3-D shapes for wing
F were generally smooth throughout the flapping cycle.
3-D Shape Characterization Also, the general bowl-shape again increased throughout
the up-flap, reversed direction at the peak angle, and de-
Once calculated images were created, the next step was to creased throughout the down-flap as the air was forced out
construct 3-D measured shapes for the six different wing of the wing. The measured images, however, more accurate-
constructions. The wings moved laterally as well as vertically, ly depict the position of the shape relative to the mechanism
as indicated by the tracking of the leading spar from the over- attachment point and show how much the rear edge of the
top perspective; the calculated images neglected this lateral wing moved to create the bowl shape. Measured imaging
motion. The out-of-plane deformations in the z-direction were was completed for the remaining five wing designs.
Fig. 16 Down-flap (a) and up-flap (b) images for the wing A design Fig. 17 Down-flap (a) and up-flap (b) images for the wing F design
The 15 images were generated for all six different wing Mechanics of Compliant Wing Designs section, the smoother
constructions. All of the up-flap and down-flap images are on the shape of the wing and the more volume captured, the better
the same scale; thus, the flapping cycles for each of the wing the flight performance. Thus, this appears to best represent the
constructions can be compared. From the images, one can general principal for associating lift and thrust results with
observe that wing F forms the smoothest shapes throughout wing designs.
the fifteen angles, and its outer edge also reaches the highest
value in the z-axis. Wings A, D, and E all had ripples and Repeatability and Sensitivity of 3-D Wing Shapes
bubbles in their shapes and do not form the smooth bowl-
shape like the images of wing F. Also, wings A and C did not Since the issues of repeatability and sensitivity in wing testing
capture large volumes like wing F. Based on these observations and construction were previously identified for lift and thrust
and the lift and thrust values generated in Characterization of measurements as well as measured volume, the sensitivity of
1570 Exp Mech (2013) 53:1561–1571
Conclusions
References 14. Mueller TJ (2001) Fixed and flapping-wing aerodynamics for mi-
cro air vehicle applications. American Institute of Aeronautics and
Astronautics, Reston
1. Pornsin-Sirirak T, Tai Y, Ho C, Keennon M (2001) Microbat: a palm- 15. Delaurier J (1993) An aerodynamic model for flapping-wing flight.
sized electrically powered Ornithopter. Proceedings of the NASA/ Aeronaut J 93:125–130
JPL workshop on biomorphic robotics, Pasadena, CA, 14–17 16. Muniappan A, Baskar V, Duriyanandhan V (2005) Lift and thrust
2. Jones KD, Bradshaw CJ, Papadopoulos J, Platzer MF (2004) Im- characteristics of flapping wing micro air vehicle (Mav). AIAA-
proved performance and control of flapping-wing propelled micro 2005-1055, Reno, Nevada
air vehicles. Proceedings of the AIAA 42nd aerospace sciences 17. Croon GCHE, de Clerq KME, Ruijsink R, Remes B, de Wagter C
meeting and exhibit, AIAA-2004-0399, Reno, Nevada (2009) Design, aerodynamics, and vision-based control of the
3. Cox A, Monopoli D, Cveticanin D, Goldfarb M, Garcia E (2002) Delfly. Int J Micro Air Veh 1(2):71–97
The development of elastodynamic components for piezoelectrical- 18. Dickinson M, Gotz K (1993) Unsteady aerodynamic performance
ly actuated flapping micro-air vehicles. J Intell Mater Syst Struct of model wings at low Reynolds numbers. J Exp Biol 174:45–64
13(9):611–615 19. Sane SP, Dickinson MH (2002) The aerodynamic effects of wing
4. Yang L-J, Hsu C-K, Ho J-Y, Feng C-K (2007) Flapping wings with rotation and a revised quasi-steady model of flapping flight. J Biol
Pvdf sensors to modify the aerodynamic forces of a micro aerial 205:1087–1096
vehicle. Sensors Actuators A Phys 139(1–2):95–103 20. Tsai B-J, Fu Y-C (2009) Design and aerodynamic analysis of a
5. Hsu C-K, Ho J-Y, Feng G-H, Shih H-M, Yang L-J (2006) A flapping flapping-wing micro aerial vehicle. Aerosp Sci Technol 13(7):383–
Mav with Pvdf-Parylene composite skin. Proceedings of the asia- 392
pacific conference of transducers and micro-nano technology 21. Hsu C-K, Evans J, Vytla S, Huang P (2010) Development of
6. Yan J, Wood RJ, Avadhanula S, Sitti M, Fearing RS (2001) Towards flapping wing micro air vehicles—design, CFD, Experiment and
flapping wing control for a micromechanical flying insect. Proceed- actual flight. 48th AIAA aerospace sciences meeting, Orlando,
ings 2001 ICRA. IEEE Int Conf Robot Autom 4:3901–3908 Florida. 1:11707–11717
7. Fenelon MAA, Furukawa T (2009) Design of an active flapping 22. Stanford B, Albertani R, Ifju P (2007) Static finite element valida-
wing mechanism and a micro aerial vehicle using a rotary actuator. tion of a flexible micro air vehicle. Exp Mech 47(2):283–294
Mech Mach Theory 45(2):137–146 23. Stanford B, Sytsma M, Albertani R, Viieru D, Shyy W, Ifju P
8. Zdunich P, Bilyk D, MacMaster M, Loewen D, DeLaurier J, (2007) Static aeroelastic model validation of membrane micro air
Kornbluh R, Low T, Stanford S, Holeman D (2007) Development vehicle wings. AIAA J 45(12):2828–2837
and testing of the mentor flapping-wing micro air vehicle. J Aircr 24. Mueller D, Gerdes JW, Gupta SK (2009) Incorporation of passive
44(5):1701–1711 wing folding in flapping wing miniature air vehicles. ASME Mech-
9. DeLuca AM, Reeder MF, Freeman J, Oi MV (2006) Flexible- and anism and Robotics Conference, San Diego
rigid-wing micro air vehicle: lift and drag comparison. J Aircr 43:2 25. Mueller D, Bruck HA, Gupta SK (2010) Measurement of thrust and
10. Gerdes JW, Gupta SK, Wilkerson S (2012) A review of bird- lift forces associated with drag of compliant flapping wing for micro
inspired flapping-wing miniature air vehicle designs. J Mech Robot air vehicles using a new test stand design. Exp Mech 50(6):725–735
4(2):021003.1–021003.11 26. Bejgerowski W, Gerdes JW, Gupta SK, Bruck HA, Wilkerson S
11. Madangopal R, Khan Z, Agrawal S (2005) Biologically inspired (2010) Design and fabrication of a multi-material compliant flap-
design of small flapping-wing air vehicles using four-bar mecha- ping wing drive mechanism for miniature air vehicles. ASME
nisms and quasi-steady aerodynamics. J Mech Des 127(4):809–817 Mechanism and Robotics Conference, Montreal, Canada
12. Bejgerowski W, Ananthanarayanan A, Mueller D, Gupta SK 27. Gerdes J (2010) Design, analysis, and testing of a flapping-wing
(2010) Integrated product and process design for a flapping-wing miniature air vehicle. Master’s thesis, University of Maryland,
drive-mechanism. J Mech Des 50:725–735 College Park
13. Bejgerowski W, Gupta SK, Bruck HA (2010) A systematic ap- 28. Peng T, Gupta SK (2007) Model and algorithms for point cloud
proach for designing multifunctional thermally conducting polymer construction using digital projection patterns. ASME J Comput Inf
structures with embedded actuators. J Mech Des 131(111009):1–8 Sci Eng 7(4):372–381