AIAA JOURNAL

Vol. 57, No. 9, September 2019

Passive Pitching of a Flapping Wing in Turning Flight

Samane Zeyghami∗
Lehigh University, Bethlehem, Pennsylvania 18015
Qiang Zhong†
University of Virginia, Charlottesville, Virginia 22903
Geng Liu‡
University of Maine, Orono, Maine 04468
and
Haibo Dong§
University of Virginia, Charlottesville, Virginia 22903
DOI: 10.2514/1.J056622
Insect wings are flexible structures that passively deform under the action of inertial and aerodynamic forces in
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

flight. Previous studies have focused on the aerodynamic and energetic merits of these deformations. Here, the effect of
torsional wing flexibility on maneuverability is investigated by modeling the dynamics of the wing pitch motion when
body yaw rotation is imposed. Analyses were carried out for different nondimensional stiffness levels, characterized
by Cauchy number and body-to-wing velocity ratios Ω∕ω. In addition, the impacts of inertial effects were evaluated
via changing mass ratio. It was demonstrated that body rotations change the balance between the aerodynamic and
elastic torque exerted about the wings’ pitching axes. This results in passive variations in the wing pitch angles that are
bilaterally asymmetric and increase linearly with the ratio of the body rotational velocity to the wing flapping velocity

The passive changes in the bilateral pitch angles induce a torque about the body’s yaw rotation, which curtails
Ω.
rotational damping effects due to the flapping motion of the wings, known as flapping counter-torque. The results
reveal that torsionally flexible wing designs could enhance maneuverability by mitigating the need for active wing
kinematic modulations during aerial maneuvers.

Nomenclature have found other benefits associated with wing flexibility, such as
b = wing span, m reduction of collision damage [9] and detection of the body angular
CL = lift coefficient velocity via sensing the changes in the structural dynamics of the
CT = thrust coefficient wing [10]. Overall, it appears that the effect of wing flexibility on
Ch = Cauchy number insect flight is broader than its mere aerodynamic merits, and more
c = wing chord, m studies need to be done to unravel its contribution to other aspects of
I = wing moment of inertia flight.
Ib = body moment of inertia Wings of insects are complex structures, consisting of a membrane
K = reduced frequency and a network of tabular veins [11]. For many insects, there exists a
M = mass ratio high-torsional-flexibility region concentrated at the wing hinge
Re = Reynolds number [6,11], which directly affects the passive pitching motion of the wing
U = wing-tip velocity, m/s during both translational phase of the flapping stroke and fast rotation
ΔA = bilateral difference in half-stroke-averaged value of any at the stroke reversal [12,13]. The network of veins across the wing
variable A surface provides another means for enhancing the torsional flexibility
Ω = yaw velocity of the body, rad∕s of the wing, causing the wing surface to twist under the action of the
ω = flapping velocity of the wing, rad∕s inertial and aerodynamic forces acting on it. Measurements on free-
flying insects have indicated that twisting can result in 10–30 deg
difference between the pitch angles of the wing root and that of the
I. Introduction wing tip [3,11,14]. Nevertheless, although twisting improves the
aerodynamic efficiency, its contribution to the magnitude of the force
E XPERIMENTAL observations and measurements have
reported flexibility of insect wings with various deformation
modes, including twisting, cambering, and spanwise bending [1–3].
is rather small [15]. Therefore, to probe how the passive deformations
of the wing affect the aerodynamics as well as the performance of the
Researchers have most commonly associated wing flexibility with insect flight, many computational studies modeled the combined
enhancing the aerodynamic performance. It has been shown that the effect of the flexibility of wing hinge and its surface as a lumped
wing deformations are beneficial to improving force production torsional spring located at the wing base [12,13,16–21].
capacity [4–6] and/or power economy [7,8]. Recent investigations Wing pitch angle and its dynamics play an important role in the
aerodynamic force generation of flapping flight. For instance, insects
Received 15 August 2017; revision received 4 January 2018; accepted for
flap their wings at high pitch angles through the air. This motion creates
publication 15 January 2018; published online 12 February 2018. Copyright a vortex at the leading edge that remains attached to the wing,
© 2018 by the American Institute of Aeronautics and Astronautics, Inc. All generating large unsteady aerodynamic forces [22]. In addition, fast
rights reserved. All requests for copying and permission to reprint should be wing flip at the stroke reversal enhances circulation on the wing,
submitted to CCC at www.copyright.com; employ the eISSN 1533-385X to providing another mechanism for aerodynamic force generation
initiate your request. See also AIAA Rights and Permissions www.aiaa.org/ [23,24]. Because of large sensitivity of both the magnitude and
randp. direction of aerodynamic force to wing pitching, many insects adjust
*Postdoctoral Researcher, Department of Mechanical Engineering and
this angle to steer and maneuver [8,17,18,25–28]. Previously, it was
Mechanics.
†
Ph.D. Student, Mechanical and Aerospace Engineering. assumed that the control of the wing kinematics, including the wing
‡
Postdoctoral Researcher, Department of Mechanical Engineering; geng pitch, is directly carried out by the flight muscles in the thorax.
[email protected]. Member AIAA. However, recent studies have suggested that the interactions of the
§
Associate Professor, Mechanical and Aerospace Engineering. flapping wing with its own unsteady flow should be considered when
3744
 ZEYGHAMI ET AL. 3745

studying the mechanics of the wing kinematics control in insects. And if so, how do the changes in wing motion influence the
For instance, investigations on the free flight of fruit flies have performance of aerial maneuvers?
indicated that the torsional flexibility of the wing hinge allows these
insects to indirectly control their wing pitch angles during aerial
maneuvers [17,18].
When an insect is in hovering or low-speed flight, the motion of the
II. Methods
bilateral wings is symmetric. However, any lateral motion A. Problem Definition
(translation or rotation) is usually induced by asymmetrically Here, we are interested only in the deceleration phase of the
varying the kinematics of the bilateral wings. Previous studies have maneuver where the active manipulation of the wing kinematics is
suggested that these changes in the wing kinematics are enforced by minimal [30,33]. Therefore, we assume that the flapping motion of
an intricate chain of actions that is controlled by the insects’ nervous the bilateral wings is symmetric. The pitching of the wing about its
system [29]. We refer to these changes as active wing kinematics rotation axis, on the other hand, may not be symmetric because it is
variations. When an animal (voluntarily) begins turning, active dictated by the combined effect of the aerodynamic, elastic, and
generation of aerodynamic torque about the desired rotation axis is inertial forces acting about the wing’s rotation axis.
inevitable. This phase of motion is referred to as the acceleration When the body is in rotation about an axis normal to the plane of
phase, where body rotational velocity increases. However, flapping, in downstroke (DS), the net rotational velocity of the outer
decelerating and stopping may not be entirely active because the wing increases, and that of the inner wing decreases. The opposite
damping forces such as friction tend to inhibit the motion [30]. occurs in upstroke (US), where the velocity of the inner wing
Particularly, a series of recent studies have suggested that the increases, and that of the outer wing decreases. Thus, the behavior of
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

dynamics of the deceleration phase in slow aerial yaw turns are the inner wing in DS is identical to that of the outer wing in US. The
dominated by a passive aerodynamic torque generation mechanism: same can be said about the behavior of the inner wing in US and outer
the so-called flapping counter-torque (FCT). They suggested that the wing in DS. Thus, to simplify the computations, we only modeled
difference in the net velocity of the bilateral wings, when the body is one wing. The difference in the dynamics of the inner and outer wings
in yaw rotation about an axis normal to the plane of flapping, induces can be obtained by analyzing DS and US of the outer wing only.
a damping torque, which acts as a breaking mechanism. This The wing is modeled as a thin cylinder with elliptical cross section.
damping torque exists even if the motion of the bilateral wings is Figure 1 shows the schematic of the wing. Wing aspect ratio, defined
symmetric. For the simplified scenario where no active control is as the ratio of the major to minor axis length of the ellipse, was kept
present, the magnitude of the damping torque is proportional to the constant. Flapping motion is defined as rotation about the y axis of the
magnitude of the body velocity, resulting in an exponential decay in body-fixed coordinate system where the origin is set at the body’s
the rate of body rotation. The predictions of this model were centerline 0.1c away from the wing root. The flapping profile is
successfully confirmed by experimental measurements on slow yaw sinusoidal with constant amplitude of 100 deg.
The dynamics of the passive wing pitching is governed by the
turns of several species of insects and birds. Nevertheless, more
combined effect of the wing morphology, structure, wing and body
recent measurements on fast yaw turns of some agile insects such as
kinematics, and flow properties. It is not immediately obvious if all
dragonflies and damselflies [26,31,32] show that the FCT model
these parameters have independent effects on the output.
alone is insufficient in explaining the dynamics of these fast aerial
Dimensional analysis was performed to combine the effect of
maneuvers.
individual parameters in the form of physically relevant nondimen-
To better understand the dynamics of fast aerial maneuvers of
sional quantities as listed follows (detailed derivation of the
flying insects and to probe how the wing flexibility plays a role in
nondimensional numbers are carried out in [12]):
enhancing maneuverability, Zeyghami and Dong [26] and Zeyghami
[32] investigated the behavior of a pair of flexible flapping

wings when they undergo yaw rotation. They employed a torsional η
f Ch; M; Re; K; Ω  (1)
spring at the wing’s rotation axis as a first-order model of the
wing’s torsional flexibility and a quasi-steady model to estimate where η is the wing pitch angle, and Ch
ρΦ2 f2 c3 b2 ∕G is the
aerodynamic force and moment. They showed that the motion of the nondimensional flexibility of the combined fluid–structure system
bilateral wings passively changes in response to the body motion. and is defined as the ratio of the fluid dynamic pressure to the
Here, we conduct a similar study in which we use an in-house high- structure elastic forces. f, c, b, and Φ are the flapping frequency,
fidelity computational fluid dynamics (CFD) tool to calculate the chord, span, and flapping amplitude, respectively. G is the torsional
aerodynamic forces and to unravel the unsteady flow features of stiffness of the spring. M
ρs t∕ρc is the mass ratio, where ρ and ρs
maneuvering flight. Similar to the aforementioned study, we model are the fluid and solid densities, and t is the wing thickness.
the wing’s torsional flexibility via a torsional spring located close to Re
ρcU∕μ, where U
2fΦb is the velocity of the wing tip.
the leading edge as a first-order approximation of the insect wing K
fc∕U is the reduced frequency. Ω 
Ω∕ω is the normalized
structure. We particularly seek to answer two questions. Does body body velocity, defined as the ratio of the body yaw velocity Ω and
rotation during aerial maneuvers affect the dynamics of wing pitch? wing flapping velocity ω
2Φf.

Fig. 1 Schematic of the model flapping wing in yaw rotation. Yaw angular velocity of the body is oriented in the positive y direction.
 3746 ZEYGHAMI ET AL.

In all simulations, we kept Reynolds number constant at 500.

Cauchy number, M, and Ω  are the inputs and vary among the
simulations. For insects, M ≈ 1 [34]. For aquatic animals, this ratio is
closer to 0.1.
The output is the wing passive pitching motion as well as the
resulting aerodynamic forces and moments. The half-stroke-
averaged force coefficients are defined as follows:

T
CT
(2)
0.5ρb3 cω  Ω2

L
CL
(3) Fig. 2 Comparison of the pitching angle among the current CFD
0.5ρb3 cω  Ω2 simulation, experimental measurement, and quasi-steady calculation [13].

where T and L are the average horizontal and vertical force in

downstroke or upstroke, respectively. The fluid force is normalized III. Results
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

by the net flapping velocity of the wing in the ground coordinate A. Flexible Flapping Wing in Yaw Rotation
system (rather than the flapping velocity in the body-fixed We start by keeping nondimensional flexibility of the wing and
coordinate). This is to separate the effect of the change in the net the mass ratio constant at Ch
0.2 and M
1. When the wing
velocity of the wing in altering the fluid forces. For the outer wing, the flaps through the air, it pitches about its rotation axis as a result of
net flapping velocity is ω  Ω in downstroke (DS) and ω − Ω in the combined effects from inertia, aerodynamics, and elastic forces.
upstroke (US). The elastic force tends to restore the orientation of the wing to its
rest orientation at η
0, where the wing surface is normal to the
B. Numerical Modeling flapping direction. The fluid force, on the other hand, tends to orient
The incompressible flow is governed by the three-dimensional the wing with the direction of the incoming velocity, where the
Navier–Stokes equations, which can be written in tensor form as surface of the wing is parallel to the flapping direction. The net
follows: effect of the inertial forces depends on the acceleration of the wing
[11,12,18]. The solid black line in Fig. 3 shows the resulting passive
wing pitch angle. Note that the pitch angle of the wing is symmetric
∂ui ∂ui ∂ui uj ∂p 1 ∂2 ui

0; 
−  (4) in DS and US. Next, we impose body rotation and repeat the
∂xi ∂t ∂xj ∂xi Re ∂xi xj numerical simulation. Figure 3 shows the passive pitch angle for
different normalized body velocities. Notice that the body rotation
in which ui (i
1; 2; 3) is the velocity components, and p is the fluid breaks the DS–US symmetry of the wing pitch angle. This is due to
pressure. Equation (4) is solved with a finite-difference-based the asymmetric change in the effective Cauchy number of the wing
Cartesian grid immersed boundary method [35]. A second-order in DS and US (mass ratio remains unchanged). When the body
central difference scheme in space is employed. Time is advanced undergoes a yaw rotation, the net incoming velocity to the wing
using a second-order-accurate fractional-step method. This method changes, altering the balance between the aerodynamic and elastic
was successfully applied in many simulations of flapping propulsion torque. In downstroke, the net velocity, and thus the dynamic
[8,24,31,36–39]. More details about the method as well as validations pressure, increases on the outer wing. The stronger fluid force
of the fluid solver can be found in our previous work [39–41]. rotates the wing more toward the incoming velocity direction,
The governing equation of the wing pitching motion is as follows: which results in decreasing pitch angle (and the geometric angle of
attack) of the wing. The opposite occurs in US, where the net
velocity, and therefore the dynamic pressure, of the wing decreases.
Ixx ω
 x  I zz − Iyyωy ωz  I xy ω
_ x − ωx ωz   I yz ω2y − ω2z  Weakening of the fluid dynamic pressure allows the restoring elastic
 I xz ω
_ z − ωx ωy 
Maero  Melastic  Mgravity (5) force to bring the wing orientation closer to its rest orientation
(vertical). Consequently, the wing pitch angle increases in US.
We will refer to the passive pitch-angle asymmetry as PPA.
where Ix x, I y y, I z z, Ix y, I x x, and I x z are elements of the wing’s
moment of inertia matrix. Maero , Melastic , and Mgravity are the torques
due to aerodynamic, elastic, and gravitational forces, respectively.
The governing equations of the fluid and the solid (the wing)
are strongly coupled via subtime step iterations, which ensure
convergence of the wing pitch angle within each time step.

C. Validation
To validate our code, a flapping wing with a torsional axis along the
spanwise direction is simulated. The passive pitching of the wing and
the flow around this wing is obtained by the aforementioned model.
The results are compared with those in Whitney and Wood [13]. The
key parameters of the validation case are as follows. The geometry of
the wing model as well as its inertial properties are taken from [13].
The flapping amplitude is set to 108 deg, and the flapping frequency
is 100 Hz. The natural frequency is fn
233.7 Hz. We compare the
simulated pitching angle of our model with that obtained by Whitney Fig. 3 Time history of the passive pitch angle of the outer wing when the
and Wood [13] through both experiments and quasi-steady model body is in yaw rotation. Different colors represent different Ω
values.
calculation in Fig. 2. Duration of downstroke is shaded in gray.
 ZEYGHAMI ET AL. 3747

0.2. The force is

Fig. 4 Time history of nondimensional lift and thrust generated by the outer wing when the body is in yaw rotation of magnitude Ω
normalized by qbc, where q
0.5ρU2 .

Figure 4 shows the lift and thrust (nondimensionalized) for the velocity. The variation of ΔCT with Ω  is relatively linear. Both the
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

same wing (Ch
0.2 and M
1) as a function of time at Ω 
0.2. magnitude of ΔCT and its rate of change with Ω  are larger for lower
Similar to the pitch angle, the force generated by the wing is Cauchy number values. The larger ΔCT values, however, are an
asymmetric in DS and US. For reference, the force generated by the artifact of the higher CT values for more rigid wings. The smaller the
wing at hover is shown by the solid red line. This asymmetry in the Cauchy number value is, the larger the relative strength of the elastic
force arises from two effects; the first effect is related to the rigid force is compared to the fluid force. This means that the wing moves
flapping motion of the wing, and the second stems from the passive with a larger angle of attack through the stroke plane and generates
response of a torsionally flexible flapping wing. The former effect larger thrust/drag force. If we normalize ΔCT by the average thrust
occurs purely due to the asymmetric change in the net incoming coefficient of the bilateral wings, C T , the lines of ΔC^ T
ΔCT ∕C T
velocity to a rigid flapping wing [30] (see dashed blue line in Fig. 4). versus Ω  collapse close to each other, as shown in Fig. 5d.
However, when a flexible wing experiences whole-body rotation, not Asymmetric change in the bilateral lift and thrust leads to the
only the incoming velocity to the wing changes asymmetrically in DS generation of roll and yaw moments, respectively. The magnitude of
and US but so does the average wing pitch angle. Thus arises the the net moment is proportional to the magnitude of asymmetry in the
difference in the force asymmetry generated by a flexible versus a bilateral forces. In the following section, we derive an estimation of
rigid wing, which is evident in Fig. 4. Note that the difference in the the net passive yaw torque generated by a pair of flexible flapping
force generated in DS and US is larger for a rigid wing when wings, when they are engaged in a whole-body rotation. In addition,
compared to a flexible wing. we compare and contrast our model against the one derived for a rigid
As was mentioned before (in Sec. II), the difference in the behavior pair of wings in [30]. Although our focus is mainly on the yaw
of the outer wing in DS and US can be interpreted as the difference in rotation and the passive yaw torque generation in the present paper, it
the behavior of the inner and outer wings during whole-body is important to note that the asymmetric changes in the bilateral pitch
rotations. Thus, the asymmetry in DS and US behavior of the wing is angles also lead to generation of a net roll moment.
representative of the asymmetry of the bilateral wings, which can
potentially lead to generation of lateral motion. In the following C. Effect of Mass Ratio
sections, the behavior of the bilateral wings during body yaw rotation To probe how PPA scales with inertial effects, we repeated our
will be analyzed based on the simulation of a single wing in the same simulations for a smaller M value of 0.1. This is closer to the mass
motion. ratio observed in aquatic animals. Figure 6 shows Δα and ΔCT as a
function of Ω.
 The trends are similar to what was observed for larger
B. Effect of Wing Flexibility mass ratio cases, but the magnitudes of the bilateral wing angle
Figure 5a shows the difference in the downstroke-averaged of attack difference and bilateral force coefficient differences are
geometric angle of attack of the bilateral wings in degrees, slightly larger for small mass ratio cases. This is not surprising
Δα
αi − αo , as a function of Ω  for different Cauchy number because, at smaller mass ratios, inertial effects are lighter, and thus
values. The geometric angle of attack is the angle between the wing dynamics of the passive wing pitch is more dominantly governed by
surface and the wing flapping velocity (stroke plane of the wing). The the interplay between the fluid and elastic forces.
magnitude of Δα is the same in DS and US, but the sign is opposite. In
DS, the inner wing has a larger angle of attack, and in US, the outer D. Effect of Passive Pitch-Angle Asymmetry on the Body Motion
wing does. The difference in the angle of attack of the bilateral wings To probe how PPA affects the body motion, here we calculate the
increases almost linearly with increase in the body velocity for a yaw torque exerted on the body due to the passive asymmetry in the
constant Cauchy number value. Both the magnitude of Δα and its rate bilateral wing pitch angles. To separate the effect of the passive wing
of change with rotational velocity is larger for more flexible wings kinematic changes from that of the body rotation only (FCT), we
(higher Cauchy number values). At M
1, Δα is not a function of make a comparison between the passive yaw torque generated by a
Cauchy number for Ch > 0.2. pair of rigid wings versus that generated by a pair of flexible wings.
Figures 5b and 5c show ΔCL and ΔCT as a function of Ω  When a pair of rigid flapping wings is engaged in a whole-body
for different Cauchy number values. ΔCT
CT i − CT o , where rotation, their net flapping velocity changes asymmetrically. This
subscripts i and o stand for inner and outer wing, respectively. Note results in a net passive yaw torque acting on the body whose
that, in the calculation of CL and CT , the force is normalized by the magnitude is proportional to the rotational velocity of the body. This
net velocity of the wing [Eq. (3)]. This definition is used to separate counteracting torque is referred to as FCT [42]:
the effect of the PPA (passive pitch asymmetry, a flexible effect) from
that of FCT (flapping counter-torque, a rigid effect) in generating 1 h i
bilateral force asymmetry. For a rigid wing, CT i
CT o and τr
ρb3 cC T l ω − Ωr 2 − ω  Ωr 2 (6)
2
CLi
CLo because the pitch angles of the bilateral wings are
identical. However, that is not the case for a flexible wing. Our results where l is the distance of the center of the action of the force from the
show that both ΔCL and ΔCT increase with the body rotational body center of mass. Ωr is the instantaneous yaw velocity of the body.
 3748 ZEYGHAMI ET AL.
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

Fig. 5 The difference in the average a) angle of attack (in degrees), b) lift coefficient, c) thrust coefficient, and d) normalized thrust coefficient as a function


of Ω.

Fig. 6
for M
0.1.

The difference in the average a) angle of attack, and b) thrust coefficient as a function of Ω

The subscript r stands for rigid. Note that, for a rigid pair of wings, Ωr
Ω0 e−4Aω∕Ib t (8)
C T
CT i
CT o . Equation (6) can be simplified as follows:
where I b is the moment of inertia of the whole-body system about
τr
−4AωΩr (7) the yaw axis of rotation, and Ω0 is the initial body yaw velocity (at the
onset of deceleration phase).
where A
1∕2ρb3 cCT l. Equation (7) indicates that the passive Similar to the case for rigid pair of flapping wings, the net flapping
yaw torque acting on the body (connected to a pair of rigid flapping velocities of the bilateral wings are asymmetric for a flexible pair as
wings) is proportional to the rotational velocity of the body, or in well. However, for a pair of flexible wings, the thrust coefficients of
other words, Ω _ r ∝ Ωr . The solution to this simple ordinary the bilateral wings are no longer identical. In downstroke, CT
differential equation is an exponential function. Thus, increases on the inner wing, and it decreases on the outer wing.
 ZEYGHAMI ET AL. 3749

The thrust coefficient of the inner and outer wing can be expressed as of the passively induced wing pitch changes, we prescribed the pitch
CT o
C T − ΔCT ∕2 and CT i
C T  ΔCT ∕2, where C T is the angle of the rigid wing as that of the flexible wing at Ω 
0 and
average thrust coefficient of the bilateral wings. Thus, the passive Ch
0.2. The Cauchy number for the flexible wing is set to 0.2
yaw torque generated by a pair of flexible wings that are engaged in as well.
whole-body rotation can be expressed as follows: The three-dimensional flow structures of both the rigid and the
flexible wing are visualized by plotting the isosurface of the maximal
    
1 ΔC^ T ΔC^ T imaginary part of complex eigenvalues of the velocity gradient
τf
ρb3 cC T l 1 ω − Ωf 2 − 1 − ω  Ωf 2 tensor, Λmax , as shown in Fig. 7. Two Λmax values of 40 and 25 are
2 2 2
visualized to show the stronger vortex core as well as the weaker
(9) surrounding vorticity. Although the major flow structures are similar,
differences in vortex evolution and shedding are identifiable.
where the subscript f stands for flexible, and ΔC^ T
ΔCT ∕C T , as
Comparing Figs. 7a–7c and 7d–7f shows that, in downstroke, the
was defined in earlier. From the previous section, we know that ΔC^ T
leading-edge vortex (LEV) of the rigid wing evolves relatively
varies linearly with Ω:
 ΔC^ T
aΩ. Thus, Eq. (9) can be rewritten as
slower. But more vorticity packs into it, and it eventually grows to
become a stronger coherent vortex, which remains attached to the
τf
−4AωΩf  AaΩf ω1  Ω
 2f  (10) wing at middownstroke. In addition, on the rigid wing, less vorticity
is shed out via the tip vortex. The footprint of this is traceable in the
 2f ≪ 1,
Because Ω plot of force versus time that is shown in Fig. 4. In downstroke, the
force peak is larger for the rigid wing, and the peak value is phase-
τf ≈ −4 − aAωΩf
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

(11) delayed when compared to the peak downstroke force of the flexible
wing. Also, the average force generated by the rigid wing in
Equation (11) has a form similar to that of Eq. (7) and can be downstroke is larger that of the flexible wing. Note that, in DS,
similarly solved for Ωf : the pitch angle of the outer (flexible) wing decreases (compared to the
hovering flight) due to the interplay between the fluid and elastic
Ωf
Ω0 e−4−aAω∕Ib t (12) forces, hence the smaller downstroke force (drag) production when
compared to the rigid wing in the same rotational flight.
Because a > 0, the rate of decay in the yaw velocity is always In upstroke, the story is reversed. Comparing the flow structure of
slower for flexible wings. The body yaw velocity half-life of a the flexible and rigid flapping wings in Fig. 8 shows that the evolution
flexible pair of wings is 4∕4 − a times larger than that of a rigid pair. of the LEVof the flexible wing is delayed when compared to the rigid
wing, and the LEV vortex of the flexible wing is stronger at
E. Effect of Body Motion on the Flow Structure of Torsionally midupstroke. The footage of this is detectable in the plot of the force
Flexible Wings versus time (Fig. 4), where the peak force (as well as the half-stroke-
Here, we compare the flow structure of a flexible versus a rigid averaged force) generated by the flexible wing in upstroke is larger
flapping wing during whole-body rotation. The nondimensional than that of the rigid wing, and the phase of the peak force is delayed
rotational velocity of the body is set at Ω

0.2. To isolate the effect when compared to the rigid wing. Similar results were found in [42].

Fig. 7 Visualization of the vortex structure of a torsionally flexible wing at a) t∕T
0.2, b) 0.3, c) 0.4, and a rigid wing at d) t∕T
0.2, e) 0.3, and f) 0.4.
 3750 ZEYGHAMI ET AL.
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

Fig. 8 Visualization of the vortex structure of a torsionally flexible wing at a) t∕T
0.65, b) 0.74, c) 0.8, and a rigid wing at d) t∕T
0.65, e) 0.74,
and f) 0.8.

IV. Conclusions 10–20 deg. For high-flapping-frequency insects, such as fruit flies
When flapping through the air, insect wings are subject to and drone flies, Ω is only about 0.05–0.1. According to the present
aerodynamic and inertial forces that influence their motion and model, at these values of Ω,  the magnitude of passive wing pitch
deformation. Through a computational study, which combines a asymmetry is only about 2–5deg, which is in the same range as the
high-fidelity fluid dynamics solver with a torsional spring structure measured wing pitch angle difference during the deceleration phase
model, we investigated how the wing pitch depends upon the body of aerial turns for these insects [18,43].
rotations. Through the present analyses, a passive mechanism has Here, the discussion was limited to the deceleration phase of the
been discovered that allows wing kinematic modulations during the aerial yaw maneuvers, where the active modulations in the wing
aerial maneuvers that were previously thought to be exclusively kinematics are minimal. Yet these passive changes in the wing
active. The passive changes in the wing kinematics act as a kinematics will also occur during the acceleration phase. In the
complementary mechanism that enhances wing pitch asymmetry acceleration phase of the maneuver, body rotational velocity, which is
during the aerial maneuvers and reduces the damping of the body initially small, increases wingbeat to wingbeat due to actively
rotational velocity. Furthermore, it was shown that the present results induced asymmetries in the motion of the bilateral wings [44]. The
and conclusions are valid for a large range of wing flexibilities and asymmetry in the wing kinematics makes the analysis of the passive
inertial properties. wing pitch modulations in the acceleration phase relatively more
To generate in-flight turns, many insects adjust the pitch angle of complicated. During this phase, the magnitude of the passive wing
their bilateral wings [18]. To tune the wing kinematics, they use the pitch asymmetry is determined by the combined effect of the
flight muscles inside the thorax. Active control of the wing motion increasing rotational velocity of the body and the asymmetrical
requires neural and mechanical feedback from the body motion as flapping velocities. Depending upon the difference in the bilateral
well as feedforward motor neurons to transport the control signal to wings’ net velocities, PPA can either enhance or reduce the
the flight muscle [29]. Therefore, any adjustment in the wing aerodynamic torque available for accelerating the body motion.
kinematics will cost time and energy. The present analysis indicated Nevertheless, the present results suggest that the effect of wing
that there is an alternative passive pathway by which changes in the flexibility should be considered throughout the whole duration of the
flight condition can affect the wing kinematics. Bypassing the active aerial maneuvers of flying insects.
control loop, the passive wing kinematic modulations are fast and The application of the present results is not limited to the
energetically efficient. While in rotation, PPA lessens the need for maneuvering flight only. The model also suggests that the commonly
active control. Moreover, because of the reduced damping of the used assumptions in the analysis of the flight stability and control of
body rotational velocity, an insect with torsionally flexible wings can the insects should be revisited. Currently, in stability analysis of the
generate larger turn angles (flight heading change) employing the insect flight, wing kinematics are assumed to remain unchanged
same amount of initial kinetic energy. In fact, the present while the motion of the body is perturbed [5,35,45–47]. The present
experimental measurements on the free-flying damselflies during fast results suggest that this assumption may not always be valid,
yaw turns (which are published in another document) show that the especially when the ratio of the body’s perturbed motion to the wing
new model, which includes the effect of wing flexibility, can flapping velocity is large, which occurs at large perturbation values
accurately predict both wing and body kinematics of these insects and/or small flapping frequencies.
during the deceleration phase of the maneuver [26,32]. Flapping It is believed that the present results have broad implications in
frequency of damselflies is only about 15–40 Hz, and thus during fast understanding insect flight dynamics, maneuverability, and wing
yaw turns, Ω  rises up to 0.4 [24,26]. At these large values of Ω,  kinematics control. For instance, the present analysis revealed that
the magnitude of passive wing pitch asymmetry is in the order of the motion of a flexible wing is tightly coupled with that of the body
 ZEYGHAMI ET AL. 3751

via interactions between the aerodynamic and elastic forces that act Interface, Vol. 7, No. 42, 2010, pp. 131–142.
on the wing. This connection is missing from the classical model of doi:10.1098/rsif.2009.0120
the insect flight dynamics in which the motion of the body affects that [15] Du, G., and Sun, M., “Effects of Unsteady Deformation of Flapping
of the wing only through the control signal that is generated after Wing on Its Aerodynamic Forces,” Applied Mathematics and
Mechanics, Vol. 29, No. 6, 2008, pp. 731–743.
processing the sensory information [29]. On another level, the doi:10.1007/s10483-008-0605-9
present results suggest that agile wing kinematic modulations may [16] Ishihara, D., Horie, T., and Denda, M., “A Two-Dimensional
not necessarily require rapid sensory information. Therefore, a higher Computational Study on the Fluid–Structure Interaction Cause of Wing
level of maneuverability may be achieved employing a simpler Pitch Changes in Dipteran Flapping Flight,” Journal of Experimental
sensory and actuation system. The outcomes of our analysis also Biology, Vol. 212, No. 1, 2009, pp. 1–10.
inform the design of micro air vehicles by unraveling the connection doi:10.1242/jeb.020404
between the wing design and maneuverability of the flight. [17] Beatus, T., and Cohen, I., “Wing-Pitch Modulation in Maneuvering
Fruit Flies Is Explained by an Interplay Between Aerodynamics and a
Torsional Spring,” Physical Review E, Vol. 92, No. 2, 2015, Paper
Acknowledgments 022712.
doi:10.1103/PhysRevE.92.022712
This work was supported by the National Science Foundation [18] Bergou, A. J., Ristroph, L., Guckenheimer, J., Cohen, I., and Wang,
(grant CEBT-1313217) and the U.S. Air Force Research Laboratory Z. J., “Fruit Flies Modulate Passive Wing Pitching to Generate
(grant FA9550-12-1-007). In-Flight Turns,” Physical Review Letters, Vol. 104, No. 14, 2010,
Paper 148101.
doi:10.1103/PhysRevLett.104.148101
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

References [19] Wan, H., Dong, H., and Huang, G. P., “Hovering Hinge-Connected
Flapping Plate with Passive Deflection,” AIAA Journal, Vol. 50, No. 9,
[1] Combes, S., and Daniel, T., “Flexural Stiffness in Insect Wings 1. 2012, pp. 2020–2027.
Scaling and the Influence of Wing Venation,” Journal of Experimental doi:10.2514/1.J051375
Biology, Vol. 206, No. 17, 2003, pp. 2979–2987. [20] Zeyghami, S., Akoz, E., and Moored, K., “Enhancing Propulsive
doi:10.1242/jeb.00523 Efficiency Through Proper Design of Bending Patterns of a Flexible
[2] Wan, H., Dong, H., and Ren, Y., “Study of Strain Energy in Deformed
Pitching Foil,” Bulletin of the American Physical Society, Vol. 61, 2016.
Insect Wings,” Dynamic Behavior of Materials, Vol. 1, May 2011,
[21] Akoz, E., Zeyghami, S., and Moored, K., “Intermittent Swimming with
pp. 323–328.
a Flexible Propulsor,” APS Meeting Abstracts, 2016.
[3] Koehler, C., Liang, Z., Gaston, Z., Wan, H., and Dong, H., “3D
[22] Birch, J. M., Dickson, W. B., and Dickinson, M. H., “Force Production
Reconstruction and Analysis of Wing Deformation in Free-Flying
and Flow Structure of the Leading Edge Vortex on Flapping Wings at
Dragonflies,” Journal of Experimental Biology, Vol. 215, No. 17, 2012,
pp. 3018–3027. High and Low Reynolds Numbers,” Journal of Experimental Biology,
doi:10.1242/jeb.069005 Vol. 207, No. 7, 2004, pp. 1063–1072.
[4] Mountcastle, A. M., and Combes, S. A., “Wing Flexibility Enhances doi:10.1242/jeb.00848
Load-Lifting Capacity in Bumblebees,” Proceedings of the Royal [23] Dickinson, M. H., Lehmann, F.-O., and Sane, S. P., “Wing Rotation and
Society B, Vol. 280, No. 1759, 2013, Paper 20130531. the Aerodynamic Basis of Insect Flight,” Science, Vol. 284, No. 5422,
[5] Zhao, L., Huang, Q., Deng, X., and Sane, S. P., “Aerodynamic Effects of 1999, pp. 1954–1960.
Flexibility in Flapping Wings,” Journal of the Royal Society Interface, doi:10.1126/science.284.5422.1954
Vol. 7, No. 44, 2010, pp. 485–497. [24] Bode-Oke, A. T., Zeyghami, S., and Dong, H., “Aerodynamics and
doi:10.1098/rsif.2009.0200 Flow Features of a Damselfly in Takeoff Flight,” Bioinspiration &
[6] Tanaka, H., Whitney, J. P., and Wood, R. J., “Effect of Flexural and Biomimetics, Vol. 12, No. 5, 2017, Paper 056006.
Torsional Wing Flexibility on Lift Generation in Hoverfly Flight,” [25] Sane, S. P., and Dickinson, M. H., “The Control of Flight Force by a
Integrative and Comparitive Biology, Vol. 51, No. 1, July 2011, Flapping Wing: Lift and Drag Production,” Journal of Experimental
pp. 142–150. Biology, Vol. 204, No. 15, 2001, pp. 2607–2626.
[7] Young, J., Walker, S. M., Bomphrey, R. J., Taylor, G. K., and Thomas, [26] Zeyghami, S., and Dong, H., “Coupling of the Wings and the Body
A. L., “Details of Insect Wing Design and Deformation Enhance Dynamics Enhances Damselfly Maneuverability,” arXiv preprint
Aerodynamic Function and Flight Efficiency,” Science, Vol. 325, arXiv:1502.06835, 2015.
No. 5947, 2009, pp. 1549–1552. [27] Alexander, D. E., “Wind Tunnel Studies of Turns by Flying
doi:10.1126/science.1175928 Dragonflies,” Journal of Experimental Biology, Vol. 122, No. 1, 1986,
[8] Li, C., Dong, H., and Liu, G., “Effects of a Dynamic Trailing-Edge Flap pp. 81–98.
on the Aerodynamic Performance and Flow Structures in Hovering [28] Zeyghami, S., Babu, N., and Dong, H., “Cicada (Tibicen Linnei) Steers
Flight,” Journal of Fluids and Structures, Vol. 58, Oct. 2015, pp. 49–65. by Force Vectoring,” Theoretical and Applied Mechanics Letters, Vol. 6,
doi:10.1016/j.jfluidstructs.2015.08.001 No. 2, 2016, pp. 107–111.
[9] Mountcastle, A. M., and Combes, S. A., “Biomechanical Strategies for doi:10.1016/j.taml.2015.12.006
Mitigating Collision Damage in Insect Wings: Structural Design Versus [29] Dickinson, M. H., Farley, C. T., Full, R. J., Koehl, M., Kram, R., and
Embedded Elastic Materials,” Journal of Experimental Biology, Lehman, S., “How Animals Move: An Integrative View,” Science,
Vol. 217, No. 7, 2014, pp. 1108–1115. Vol. 288, No. 5463, 2000, pp. 100–106.
doi:10.1242/jeb.092916 doi:10.1126/science.288.5463.100
[10] Eberle, A., Dickerson, B., Reinhall, P., and Daniel, T., “A New Twist on [30] Hedrick, T. L., Cheng, B., and Deng, X., “Wingbeat Time and the
Gyroscopic Sensing: Body Rotations Lead to Torsion in Flapping, Scaling of Passive Rotational Damping in Flapping Flight,” Science,
Flexing Insect Wings,” Journal of the Royal Society Interface, Vol. 12, Vol. 324, No. 5924, 2009, pp. 252–255.
No. 104, 2015, Paper 20141088. doi:10.1126/science.1168431
doi:10.1098/rsif.2014.1088 [31] Zeyghami, S., Bode-Oke, A. T., and Dong, H., “Quantification of Wing
[11] Ennos, A. R., “The Importance of Torsion in the Design of Insect and Body Kinematics in Connection to Torque Generation During
Wings,” Journal of Experimental Biology, Vol. 140, No. 1, 1988, Damselfly Yaw Turn,” Science China Physics, Mechanics & Astronomy,
pp. 137–160. Vol. 60, No. 1, 2017, Paper 014711.
[12] Ishihara, D., Yamashita, Y., Horie, T., Yoshida, S., and Niho, T., “Passive [32] Zeyghami, S., “Wing in the Loop; Integrating the Wing into Dynamics
Maintenance of High Angle of Attack and Its Lift Generation During of Insect Flight,” Ph.D. Dissertation, Univ. of Virginia, Charlottesville,
Flapping Translation in Crane Fly Wing,” Journal of Experimental VA, 2016.
Biology, Vol. 212, No. 23, 2009, pp. 3882–3891. [33] Cheng, B., and Deng, X., “Translational and Rotational Damping of
doi:10.1242/jeb.030684 Flapping Flight and Its Dynamics and Stability at Hovering,” IEEE
[13] Whitney, J. P., and Wood, R. J., “Aeromechanics of Passive Rotation in Transactions on Robotics, Vol. 27, No. 5, 2011, pp. 849–864.
Flapping Flight,” Journal of Fluid Mechanics, Vol. 660, Oct. 2010, doi:10.1109/TRO.2011.2156170
pp. 197–220. [34] Dai, H., Luo, H., and Doyle, J. F., “Dynamic Pitching of an Elastic
doi:10.1017/S002211201000265X Rectangular Wing in Hovering Motion,” Journal of Fluid Mechanics,
[14] Walker, S. M., Thomas, A. L., and Taylor, G. K., “Deformable Wing Vol. 693, Feb. 2012, pp. 473–499.
Kinematics in Free-Flying Hoverflies,” Journal of the Royal Society doi:10.1017/jfm.2011.543
 3752 ZEYGHAMI ET AL.

[35] Taylor, G., and Thomas, A., “Animal Flight Dynamics 2. Longitudinal Vol. 242, June 2013, pp. 946–990.
Stability in Flapping Flight,” Journal of Theoretical Biology, Vol. 214, doi:10.1016/j.jcp.2013.01.014
No. 3, 2002, pp. 351–370. [42] Zhong, Q., Liu, G., Ren, Y., and Dong, H., “On the Passive Pitching
doi:10.1006/jtbi.2001.2470 Mechanism in Turning Flapping Flights Using a Torsional Spring
[36] Liu, G., Dong, H., and Li, C., “Vortex Dynamics and New Lift Model,” 47th AIAA Fluid Dynamics Conference, AIAA Paper 2017-
Enhancement Mechanism of Wing–Body Interaction in Insect Forward 3817, 2017.
Flight,” Journal of Fluid Mechanics, Vol. 795, May 2016, pp. 634–651. [43] Zhang, Y., and Sun, M., “Wing Kinematics Measurement and
doi:10.1017/jfm.2016.175 Aerodynamics of Free-Flight Maneuvers in Drone-Flies,” Acta
[37] Liu, G., Ren, Y., Zhu, J., Bart-Smith, H., and Dong, H., “Thrust Mechanica Sinica, Vol. 26, No. 3, 2010, pp. 371–382.
Producing Mechanisms in Ray-Inspired Underwater Vehicle Propulsion,” doi:10.1007/s10409-010-0339-2
Theoretical and Applied Mechanics Letters, Vol. 5, No. 1, 2015, [44] Fry, S. N., Sayaman, R., and Dickinson, M. H., “The Aerodynamics of
pp. 54–57. Free-Flight Maneuvers in Drosophila,” Science, Vol. 300, No. 5618,
doi:10.1016/j.taml.2014.12.004 2003, pp. 495–498.
[38] Fish, F. E., Schreiber, C. M., Moored, K. W., Liu, G., Dong, H., and Bart- doi:10.1126/science.1081944
Smith, H., “Hydrodynamic Performance of Aquatic Flapping: [45] Xu, N., and Sun, M., “Lateral Dynamic Flight Stability of a Model
Efficiency of Underwater Flight in the Manta,” Aerospace, Vol. 3, Bumblebee in Hovering and Forward Flight,” Journal of Theoretical
No. 3, 2016, p. 20. Biology, Vol. 319, Feb. 2013, pp. 102–115.
doi:10.3390/aerospace3030020 doi:10.1016/j.jtbi.2012.11.033
[39] Wan, H., Dong, H., and Gai, K., “Computational Investigation of Cicada [46] Sun, M., and Xiong, Y., “Dynamic Flight Stability of a Hovering
Aerodynamics in Forward Flight,” Journal of the Royal Society Bumblebee,” Journal of Experimental Biology, Vol. 208, No. 3, 2005,
Interface, Vol. 12, No. 102, 2015, Paper 20141116. pp. 447–459.
[40] Mittal, R., Dong, H., Bozkurttas, M., Najjar, F., Vargas, A., and von doi:10.1242/jeb.01407
Downloaded by BEIHANG UNIVERSITY on January 11, 2021 | http://arc.aiaa.org | DOI: 10.2514/1.J056622

Loebbecke, A., “A Versatile Sharp Interface Immersed Boundary [47] Cheng, B., Deng, X., and Hedrick, T. L., “The Mechanics and Control of
Method for Incompressible Flows with Complex Boundaries,” Journal Pitching Manoeuvres in a Freely Flying Hawkmoth (Manduca Sexta),”
of Computational Physics, Vol. 227, No. 10, 2008, pp. 4825–4852. Journal of Experimental Biology, Vol. 214, No. 24, 2011, pp. 4092–4106.
doi:10.1016/j.jcp.2008.01.028 doi:10.1242/jeb.062760
[41] Chang, C.-H., Deng, X., and Theofanous, T. G., “Direct Numerical
Simulation of Interfacial Instabilities: A Consistent, Conservative, All- F. N. Coton
Speed, Sharp-Interface Method,” Journal of Computational Physics, Associate Editor