Turning on a dime: Asymmetric vortex formation in hummingbird maneuvering flight

Abstract

A pure yaw turn of a calliope hummingbird is investigated through high-speed photogrammetry, three-dimensional surface reconstruction, and computational fluid dynamics simulations. We took high-speed videos of freely flying hummingbirds using three synchronized Photron high-speed cameras. The hummingbird body and wing motions were then reconstructed using a joint-based hierarchical subdivision surface method [1,2]. As an example, Fig. 1(a) shows a high-speed image of the maneuvering hummingbird and Fig. 1(b) presents the corresponding reconstructed model. There are in total six strokes of this flight and they can be divided into three flying phases: a starting phase (first stroke), a maneuvering phase (second to fourth strokes), and a recovering phase (fifth and sixth strokes).A Cartesian grid based immersed boundary Navier-Stokes solver is used to simulate the unsteady flows around the body and wings of the bird [3,4]. We first look at the near wake structures by cutting two-dimensional flow slices along the hummingbird wing span. As shown in Fig. 1(c), leading edge vortex (LEV) structures at five flow slices can be observed for each wing. The contours of the LEV represent the normalized spanwise vorticity, which reveals the LEV strength. From the contour we observe that the LEV strength of the inner wing is greater than that of the outer wing. We further quantify the LEV circulation in the maneuvering phase. The results show that the inner wing produces 21% and 26% more average LEV circulation than the outer wing does during the downstrokes and upstrokes, respectively.The three-dimensional wake structures are identified by plotting the isosurface of Q criterion of the flow. As shown in Fig. 1(d), asymmetric wake structures between the inner and outer sides of the hummingbird body can be identified. Two new vortices, a shed trailing edge vortex (STEV) and a shed leading edge vortex (SLEV), can be identified at the outer side of the hummingbird body below the LEV. The LEV, tip vortex (TV), and STEV are connected to form a vortex ring near the outer wing tip region. Also, the LEV, SLEV, trailing edge vortex (TEV), and root vortex (RV) are connected to form another vortex ring. The two vortex rings form a unique dual-ring vortex structure at the outer side of the hummingbird body. In contrast, the LEV, TV, TEV, and RV are connected to form a single-ring vortex structure at the inner side of the hummingbird body.In order to better understand the wake topology and its effect on the aerodynamic performance, nondimensional pressure is plotted to indicate the low-pressure regions behind the wings of the

Quantifying the dynamic wing morphing of hovering hummingbird

“…25-50 • at mid-downstroke and −75 to −20 • at mid-upstroke; see figure 17). Together with the thin leading edge at the wing distal portion (because the p10 feather vane is constituting the leading edge there, figure 12), it is highly likely the LEV is generated during both mid-strokes on the distal portion of the wing, as observed in previous studies [5,6,46].…”

A computational investigation of lift generation and power expenditure of Pratt’s roundleaf bat (Hipposideros pratti) in forward flight

“…Numerical analysis is adding a new dimension to the study of unsteady aerodynamic flows. Recently, significant progress has been made towards the simulation of animal flight aerodynamics—for example, dragonflies [33,34], cicadas [35,36], and hummingbirds [37,38]. One feature all these animals have in common is their relatively stiff wings in comparison to bats.…”

Extracting signature responses from respiratory flows: Low‐dimensional analyses on Direct Numerical Simulation‐predicted wakes of a flapping uvula

“…Three pressure probes are marked as red dots, located, respectively: (1) in the back wall of the pharynx, (2) immediately downstream of the uvula, and (3) 12 mm downstream of the uvula tip. Six models will be considered from the combination of K1-2 and m1-3 (K1m1, K1m2, K1m3, K2m1, K2m2, and K2m3) insects, 48,49 birds, 50,51 and fishes. 52,53 More details of this solver can be viewed in our previous works.…”

“…An in‐house DNS solver 47 based on Cartesian‐grid finite‐difference immersed‐boundary method was implemented in this study (Figure 2A). This solver has been well tested in the flapping propulsion simulations for insects, 48,49 birds, 50,51 and fishes 52,53 . More details of this solver can be viewed in our previous works 53,54 .…”