Session X02: Biological Fluid Dynamics: Flying (10:45am - 11:30am CST)

High-fidelity simulations of hummingbird escape maneuver

Hummingbirds are extremely agile among the flying animals. In this study, we reconstructed the full-body kinematics of the escape maneuver from high-speed videos of a hovering hummingbird when it was startled. The reconstruction includes both wings, head, trunk, and the tail motion. The kinematics was then incorporated into the 3D CFD simulation using a parallel immersed-boundary method. From the flow simulation and force analyses, we will identify the critical forces that allow the bird to perform such a rapid maneuver sequence, including pitching backward, rolling, and accelerating, etc. In addition, we will discuss the unsteady aerodynamics involved in the two wings. Project is funded by the ONR.   

A CFD simulation of a flying snake gliding with body undulations

Flying snakes are the only species of snakes on Earth capable of gliding, taking advantage of fluid dynamic principles to leap from point to point among the trees. Because they undulate in the air, their unsteady vortex dynamics are critical to understanding their aerodynamics. However, no detailed flow field information can be obtained due to the limitations in experimental flow visualization techniques. In this study, a combined motion capture technology and numerical study has been conducted to study the fluid dynamics of a flying snake gliding. With the high-speed video of the snake gliding shot, its body and kinematic model was reconstructed for computational fluid dynamic (CFD) simulation. An immersed-boundary-method (IBM)-based direct numerical simulation (DNS) flow solver along with adaptive mesh refinement (AMR) was used to simulate the corresponding unsteady flows around the snake body and its flying path. Analysis has been performed on vortex dynamics and spanwise velocity feature. Results from this study are expected to bring more comprehensive understanding of flying snakes' gliding pattern and its flow field and further provide insights into the design and optimization of bio-inspired robots from an aerodynamic perspective.   

How the flapping wing kinematics and flight trajectories modulate the odor plume structure in the odor tracking flight of fruit flies?

In nature, many insects rely on their olfactory system for detecting food sources, prey, and mates. They can sense odorant plumes emitting from sources of their interest, use their highly efficient flapping-wing mechanism to follow odor plumes, and track down odor sources. The odor-tracking process typically consists of two distinct behaviors: surging upwind and zigzagging crosswind. In this study, a fully coupled three-way flight simulator is developed, which solves the 3D Navier-Stokes equations, tightly coupled with equations of motion for the passive flapping wings, and the advection-diffusion equations for the odor concentration. This simulator will be applied to investigate the unsteady flow field and the odorant transport phenomena of a fruit fly model in both surging upwind and crosswind casting. We hypothesize that the unsteady flow generated during flapping flight would perturb the odor plumes structures and significantly impact the mass transport of odorant to the olfactory receptors (i.e., antennae). Our simulation results will provide new insights into the mechanism of how fruit flies perceive odor landscape and inspire the future design of odor-guided unmanned robotic flyers for surveillance and detection missions.   

Unsteady flow and force control for flies landing upside down on a ceiling

The process landing upside down on a ceiling (i.e., inverted landing) for a fly is a common, yet complex aerodynamic feat. It is known that these inverted landing maneuvers require a sequence of well-coordinated behavioral modules of the body. However, the wing kinematics of these physical maneuvers remain largely unknown, and the fluid dynamic principles underlying this sophisticated behavior are still out of our grasp. In this work, we turned a high-speed photogrammetry of the inverted landing of a blue bottle fly (Calliphra vomitoria) into a 3D surface reconstruction. The reconstructed data was used to investigate the wing kinematics of a fly during an inverted landing. High fidelity simulations were then carried out in order to understand vortex formation in both near-field and far-field of flapping wings and examine the associated aerodynamic performance. A Cartesian grid based sharp interface immersed boundary solver was used to handle such unsteady flow simulations in all their complexity. Our simulation results of aerodynamic forces indicated that, as the fly approaching the ceiling, it reorients the wing stroke plane more vertically and produces minimum force along the horizontal direction.   

Wing kinematics and unsteady aerodynamics of hawkmoth in hovering and forward flight

The hawkmoth (Manduca sexta) is able to control its flight speed by modifying the flapping motion of its wings. As a result, the hawkmoth's wing kinematics, lift/drag force generation, and power requirements vary across different flight speeds. The goal of this work aims to compare the aerodynamics involved in hawkmoth hovering to those in forward flight. High-speed video recordings and 3D surface reconstruction were used to capture a hawkmoth's wing kinematics when hovering at 0m/s and at forward flight speeds of 2m/s, 3m/s, and 4m/s. Following reconstruction, the insect model was imposed in an in-house immersed-boundary-method based computational fluid dynamics (CFD) solver. The CFD solver provided a quantitative measure of the force generation, power requirements, and complex vortex structures generated during sustained flight. These results enable the analysis of certain trends in how hawkmoths adjust flapping kinematics and its associated unsteady aerodynamics across different flight speeds.   

Parameter identification for the mass-spring model of Calliphora wings using optimization with genetic algorithm and applications in flapping insect flight

The secret to the spectacular flight capabilities of insects is hidden in their wings which can undergo significant deformation during flight. In the current work, we present detailed numerical simulations of fluid-structure interaction (FSI) modeling a tethered flapping fly with such deformable wings in both laminar and turbulent flows. The wing dynamics is taken into account by using a mass-spring model. The method is chosen because of its simplicity and computational efficiency. However, setting the optimal stiffness parameters for the mass-spring system plays a crucial role in obtaining a realistic wing model. To overcome this challenge, we propose an approach based on the covariance matrix adaptation genetic strategy (CMA-ES). These parameters are optimized using CMA-ES by comparing the static deformations of the wing model with the referenced deformations measured from experiments. The flexible wing model with the optimized parameters will be then coupled with the fly body for FSI simulation. The code, designed for running on massively parallel supercomputers, allows us to have some insights about the impact of wing flexibility on the aerodynamic performance of winged insects.   

Trans-continental migration of dragonfly Pantala Flavescens between India and Africa: Energetics and Role of wind

The dragonfly species \textit{Pantala flavescens }migrates between India and Africa, covering a distance of around 14000-18000 km, crossing the Indian ocean twice, which is very extraordinary for a dragonfly or any flier. The route followed in this migration is India-Maldives-Seychelles-Africa-India, comprising of four legs. For such a migration, flight energetics and role of wind are most important factors. Computation models were developed for energetics, optimal time and route estimation. Energetics estimation shows, a P.flavescens can fly for 90hrs, covering a distance of 1400km; Without wind, time required for the completion of migration is more than 90hrs for all the legs. Our results show that with wind assistance, the time taken is well within 90hrs for all the legs. Also, P. flavescens detour from the virtual direct line connecting the two points of a leg in the sense of geodesic, and it follows the wind, which is expected based on the wind compensation capability of dragonflies. The results clearly show that wind assistance is vital if P.flavescens has to complete the migration. Pantala flavescens is able to achieve this great feat with the help of winds and stand in league of birds as far as migration flight is concerned.   

Effect of cruising speed on the flight performance of a dragonfly wing

The flight dynamics of a dragonfly is significantly affected by its corrugated wing. Two-dimensional numerical simulations have been performed to study the flow around a corrugated wing pitching about quarter chord point. The objective is to understand the effect of variation in Reynolds number (by varying cruising speed) on the bio-inspired corrugated wing, thereby unravelling the flight dynamics of a dragonfly at various cruising speed. Numerical simulations are conducted over a wide range of Reynolds numbers, Re $=$ 200 - 3000, for~various pitching frequencies and pitching amplitude of $5^{o}$. The results for corrugated wing are compared against a smooth NACA (National Advisory Committee for Aeronautics)0012 wing. Over the studied Reynolds number range, corrugated wing could not outperform the NACA0012 wing in terms of thrust generation and net thrust has been observed for both the wings. At Re $=$ 200, the force coefficient for corrugated wing is found to be significantly lower than that for NACA0012. The present study helps understand the influence of cruising speed in the flight dynamics of dragonflies.   

Towards hearing mosquito-borne diseases

Mosquito bites are responsible for the transmission of deadly pathogens among humans, including Plasmodium parasites, which cause Malaria as well as viruses such as Zika, leading to deaths of more than 0.5 million people every year. Eastern equine encephalitis (EEE) and West Nile virus (WNV) are the latest mosquito-borne threats to appear in the northeastern United States. Insects flap their wings to fly, generating sound due to the scattering of turbulence at the trailing edge. Interestingly, insects utilize this unavoidable consequence for mating purposes, as the sound generated by the wingbeat is usually different for males and females of the same species. Researchers have taken advantage of the flapping sounds of natural fliers for numerous purposes, including the design of quiet drones. Possibly, one of the most significant applications of flight tone detection is to identify disease-carrying insects using their noise footprints, which can be used to build an early warning system to prevent mosquito-borne diseases. In order to fulfill this goal, we apply deep learning algorithms to classify insect species utilizing the data of mosquito wingbeat sounds. Moreover, the connection between the flight tones will be discussed with the flying mechanisms.   

Effects of added mass and wing-wake interaction on the aerodynamic performance of a hovering mosquito wing.

We investigate the aerodynamic force-production mechanisms of a hovering mosquito wing. In particular, the effect of added mass and wing-wake interaction at~an~early phase of each stroke is examined through the kinematic control of sweeping and deviation motions. The results show that the added-mass force has a significant contribution to the total aerodynamic force during and near the stroke reversal. In fact, the mosquito takes an advantage of added-mass effect to produce the additional lift force to compensate the small sweeping amplitude. Also, this study shows that the wing-wake interaction affects the aerodynamic performance of the wing in two different ways. Firstly, wake from the previous stroke induces the downwash velocity to reduce the resultant aerodynamic force. Secondly, LEV formed in the previous stroke slides along the pressure side and reattach on the suction side of the wing. This results in an attached TEV formed in the successive stroke which contributes to the positive lift force right after the stroke reversal. Furthermore, it is observed that the deviation motion only affects the formation of TEV while the wake effect is not affected by the deviation motion.   

The study of a dragonfly flight performance at various pitching angles of its wing

The flight of a dragonfly is affected by its corrugated wing cross-section. The flow around a corrugated foil inspired from a dragonfly wing, pitching about quarter chord point has been studied through numerical simulations. The objective is to understand the effect of variation in pitching angle on the corrugated foil, thereby unravelling the secret of dragonfly flight at various pitching angles of its wing. Numerical simulations are performed over a range of pitching angles, $\theta =2^{o}-10^{o}$ for a reduced frequency of 8.73, at a chord-based Reynolds number of 2000. The results for corrugated foil are compared with a smooth NACA (National Advisory Committee for Aeronautics)0012 foil. Over the studied pitching angle range for the corrugated foil, at $\theta =2^{0}$ net drag on the foil is observed which transits into a net thrust on the foil at $\theta =5^{o}$ and thereafter a monotonic increase in thrust is observed up to the largest pitching angle. NACA00012 foil is observed to perform better in terms of thrust compared to corrugated foil. This study helps understand the effect of pitching angle in the flight dynamics of dragonflies.   

Wing Flapping by a Tiny Parasitoid Wasp

Tiny insects such as thrips and parasitoid wasps are important agricultural pests and also serve as bio-control agents for other insect pests. The flight of these insects may be important in their dispersal across agricultural fields, but their flight capabilities and aerodynamics are not well understood. These mm-scale insects flap their wings at beat frequencies of several hundred Hz and use unsteady aerodynamic interactions between the wings (e.g. the clap-and-fling maneuver) to generate lift in order to fly, but the aerodynamics of such low Reynolds number flapping is not well studied. Here we investigate the flapping behavior of the parasitoid wasp \textit{Trichogramma minutum} which we have observed to flap its wings using the clap-and-fling maneuver while at rest on a surface. This behavior may actively draw odorants to the insect and thus enhance olfactory navigation. The wasp has a body length of approximately 0.46 mm and teardrop-shaped fringed wings with length and maximum chord width of 0.43 mm and 0.22 mm, respectively. We provide visualization of this behavior using an ultra-high speed brightfield microscopy system and describe efforts to measure the flow fields generated by Trichogramma wing flapping while at rest on a surface and while flying.   

Vortex structure comparison between experimental and computational studies on a hovering hawkmoth.

In this work, the first full vortex structure comparison between experiments and computations is carried out. From a high-speed \textit{Schlieren} photography on a freely flying hawkmoth, salient flow structures were successfully visualized and captured. In the down-stroke, a vortex loop structure was created on each wing. It was formed by the leading-edge vortex, the tip vortex, the starting/stopping vortex, and the root vortex. In the up-stroke, after wing supination, the vortex loop was shed into the wake, while two significant tip vortices were created from the tips of fore- and hind-wings with ends connecting to the previously shed vortex loop. Concurrently, on each wing, a root vortex was created and linked to the shed vortex loop. The vortex structures are late reconstructed three-dimensionally utilizing the Direct Linear Transformation. On the other hand, an immersed boundary method based numeric computation is conducted on a hovering hawkmoth and resolves the same vortical flow phenomena on hovering hawkmoths. Meanwhile, the vortex structures are compared quantitatively from estimating the vortex loop areas formed in down- and up-strokes, showing good agreements between the experiments and numeric simulations.   

Flexible clap and peel in the smallest insects

Of the insects that have been filmed in flight, those that are 1 mm in length or less often clap their wings together at the end of each upstroke and fling or peel them apart at the beginning of each downstroke. This motion both enhances the lift forces generated during flight and also dramatically increases the drag required to fling the wings apart. Since the horizontal component of the forces acting on each wing at the end of the upstroke and beginning of the downstroke nearly cancel, the horizontal force does not contribute to thrust, lowering the aerodynamic efficiency of flight. In this presentation, a 3D parallelized and adaptive implementation of the immersed boundary method is used to numerically simulate clap and peel in rigid and flexible wings. We find that the drag forces generated during peel in 3D are substantially lower than those generated in 2D due to the strong flow normal to the root-to-tip axis. Drag is further reduced when flexibility is incorporated. In some instances, the net lift forces generated are also improved relative to the rigid wing case.   

Three-dimensional flapping flight with bristled wings at low Reynolds numbers

Miniature flying insects with body lengths under 2 mm, such as thrips and several parasitoid wasps, are often observed to possess bristled wings and use wing-wing interaction (clap-and-fling) during free flight. Previous studies using 3D wingbeat kinematics have primarily focused on examining force generation by solid (non-bristled) wings across varying Reynolds number (Re). Our previous 2D clap-and-fling studies have shown that bristled wings augment lift-over-drag ratio at Re relevant to the tiny insect flight (Re$=$10). This study aims to evaluate if the aerodynamic benefits of bristled wings are also observed when using 3D wingbeat kinematics. A dynamically scaled robotic model capable of replicating realistic 3D wingbeat kinematics during hovering was developed. We comparatively examine the flow structures generated by solid and bristled wings at Re ranging from 1 to 120, for prescribed wing kinematics of both fruit flies and thrips. The implications of observed flow structures on force generation will be discussed.   

Numerical study of a passively deployable flap for aerodynamic flow control

Birds can perform low-speed maneuvers at post-stall angles of attack (AoAs), owing in part to covert feathers -- a set of self-actuating feathers located on the suction surface of the wings. During unsteady flow separation at large AoAs, these feathers protrude into the flow and provide lift enhancements, for reasons that are still not understood. To facilitate the use of covert-feather-inspired designs in bio-inspired aerial vehicles, and to enable plausible hypotheses for the utility of these feathers in biological flight, we investigate a model system in which a passively deployable, torsionally hinged flap is mounted on the suction surface of a stationary airfoil. We perform high-fidelity nonlinear simulations to quantify the effect of flap moment of inertia, torsional stiffness, and chordwise location on aerodynamic performance. Lift improvements as high as 30\% relative to the baseline flapless case are observed when the flap is located near the mid-chord. We relate lift benefits for the optimal flap parameters to their effect on formation, strength, and interaction of key vortical structures, and provide governing dimensionless parameters for the system.   

3D Reduced-order modeling of flapping flight with heavy and highly articulated wings

A three-dimensional reduced-order modeling technique is described for modelling and simulating inertial and aerodynamic forces associated with freely-flying flapping animals and robots. The model is then applied to previously-measured flights performed by different individuals of lesser-nosed dog-faced fruit bat. Both the inertial forces and torques of the flapping and folding of the wing, as well as the ensuing aerodynamic forces are considered. Quasi-steady Blade Element Momentum theory (BEMT) is used to model the aerodynamic forces on each segment of the highly articulated wing, and the center of mass and moment of inertia of the wing are computed and incorporated into the Lagrangian equation of motion for the overall animal dynamics. Comparisons between the predicted and observed body motions indicate that both thrust and weight support are captured well, although the model shows high sensitivity in the pitch axis. The origins of this sensitivity are also discussed.   

Passive aeroelastic deflection of avian primary feathers

Bird feathers are complex structures that passively deflect in response to aerodynamic loading via spanwise bending, twisting, and sweeping. These deflections are hypothesized to optimize flight performance, but this has not yet been tested. We measured deflection of isolated feathers in a wind tunnel to explore how flexibility altered aerodynamic forces in emulated gliding flight. We predicted that 1) feathers would deflect under aerodynamic load, 2) bending would result in lateral redirection of force, and 3) twisting would alter spanwise $\alpha $ ``washout'' and delay the onset of stall. We found that bending resulted in the redirection of lateral forces toward the base of the feather on the order of 10 percent of total lift. In comparison to the airfoil which stalled at $\alpha =$13.5 deg, all feathers continued to increase lift production with increasing angle of attack to the limit of our range of measurements ($\alpha =$27.5 deg). These results demonstrate that feather flexibility may provide passive roll stability and delay stall by twisting to reduce local $\alpha $ at the feather tip. Our findings are the first to measure forces due to feather deflection under aerodynamic loading and can inform future models of avian flight as well as biomimetic morphing-wing technology.   

Turbulence explains the accelerations of an eagle in natural flight

Soaring birds travel hundreds of miles without flapping their wings, and previous work has shown birds' ability to~utilize large-scale atmospheric flow structures such as thermal and orographic updrafts to enhance flight. However, it is~unclear what role turbulence plays in avian flight behavior. We analyzed acceleration and GPS data from a~golden eagle (\textit{Aquila chrysaetos}) and found that the bird's accelerations can be entirely explained by linear interactions with turbulence within an~interval of timescales between about 0.5 and 10 seconds. We isolated soaring flight from other behaviors by finding~patterns in the accelerations and by adapting known methods for behavior classification. From the position data, we approximated the wind speeds experienced by the golden eagle, which were between 2 and 13 m/s.~We found that the probability distribution function and the spectrum of the eagle's accelerations are non-Gaussian and~resemble those for lightly inertial particles in turbulence. Additionally, we find that higher wind speeds are associated~with larger accelerations, corresponding to a lifting of the spectra consistent with the increasing strength of turbulence~in higher winds.~   

Vortex Formation Time of Animal Flight

A recurring feature of an animal's flight is a roll-up of free shear layers into swirling vortical structures. During these roll-up processes, the animals experience unsteady aerodynamic forces associated with the forming vortex that can enhance their propulsive magnitude and efficiency. Thus, a dimensionless number describing the vortex formation might prove useful in describing the similarity in animal flight across different taxa. The growth of the vortex is influenced by free stream velocity, wing kinematics, and wing morphology. Grounded in the vortex formation process, we defined various dimensionless times that arise from these parameters. Using the existing data in the literature, we compared dimensionless times to explore the possible uniformity across taxa and flight conditions ranging from hovering to cruising.   

Study of Multiple Covert-Inspired Lift-Enhancing Flaps

During high-angle attack operations, such as flying through gust or landing on a perch, a system of feathers known as the coverts is deployed passively. Coverts are referred to as nature's aeroelastic flow control devices. There are usually multiple rows of coverts on the suction side of a bird's wing. In this study, multiple covert-inspired flap rows were tested in the wind tunnel on a two-dimensional wing section at Re$=$2e5. Experiments were conducted at angles of attack (AoA) ranging from -4 deg to 46 deg. The multi-flap design included five flaps mounted on the suction surface of the NACA 2414 airfoil. During the wind tunnel tests, the flaps can either be activated to freely-move or be deactivated to stay closed. Lift improvements of up to 41{\%} in the post-stall angle of attack regime were shown at AoA$=$20 deg for the five-flap configuration. A data-driven model was developed to show that each flap's lift improvement can be superimposed. The lift improvement due to the five-flap configuration is approximately equal to the sum of the lift improvements from each flap. The freely-moving flap angles were recorded during testing. Results presented in this study can help design new flow control devices for UAVs and shed insight into the role of the coverts during bird flight.   

Towards Mission Adaptability of Small UAVs: A Leading-Edge Alula-Inspired Device (LEAD)

The Leading-Edge Alula-Inspired Device (LEAD) is a bioinspired flow control device developed based on the alula feather structure on a bird wing. In bird flight, the alula only deploys during complex maneuvers such as take-off, landing, and perching. On an engineered wing, the addition of an alula-inspired device, such as the LEAD, enhances lift and mitigates stall at high angles of attack. Here, we explore an alula-inspired leading-edge device installed on a high-lift airfoil and a moderate aspect ratio wing. Wind tunnel experiments are conducted at post-stall and deep-stall angles of at Reynolds numbers of 100,000. Experimental results including integrated force measurements and hotwire anemometry, and PIV are discussed. We examine the distinct effects of the geometric parameters of a LEAD on the aerodynamic performance of both an airfoil and an finite wing. Results show that these lift improvements are more sensitive to the LEAD relative angle of attack and root location than to the LEAD tip deflection angle. The LEAD affects the airflow in two fundamental ways. First, it increases the capability of the wing to maintain higher pressure gradients by modifying the near-wall flow close to the leading-edge. Second, it generates tip vortices that modify the structure of the turbulence on the upper-surface of the wing, delaying flow separation