Determination of Spatial Fidelity Required to Accurately Mimic the Flight Dynamics of a Bat

Bioinspiration & Biomimetics PAPER Determination of spatial fidelity required to accurately mimic the flight dynamics of a bat To cite this article: Peter Windes et al 2019 Bioinspir. Biomim. 14 066011 View the article online for updates and enhancements. This content was downloaded from IP address 18.218.56.169 on 23/10/2019 at 17:39 IOP Bioinspir. Biomim. 14 (2019) 066011 https://doi.org/10.1088/1748-3190/ab3e2a Bioinspiration & Biomimetics Bioinspir. Biomim. 14 PAPER 2019 Determination of spatial fidelity required to accurately mimic 2019 IOP Publishing Ltd RECEIVED © 13 May 2019 the flight dynamics of a bat REVISED 4 August 2019 BBIICI Peter Windes1 , Danesh K Tafti1,3 and Rolf M ller1,2 ACCEPTED FOR PUBLICATION ü 23 August 2019 1 Department of Mechanical Engineering, Virginia Tech, Blacksburg, VA, United States of America 2 Shandong University Virginia Tech International Laboratory, Jinan, People s Republic of China 066011 PUBLISHED ’ 24 September 2019 3 Author to whom any correspondence should be addressed. E-mail: [email protected] P Windes et al Keywords: bat flight, unsteady aerodynamics, motion capture, computational fluid dynamics, flapping flight, animal flight Supplementary material for this article is available online Abstract Bats possess unique flight capabilities enabled by their wing morphology. While the articulated bone structure and flexible membrane constituting the wing are known to play a critical role in BB aerodynamic performance, the relationship has never been robustly quantified. Characterization of the sensitivity between precise wing contour and aerodynamic performance is important when 10.1088/1748-3190/ab3e2a designing a biomimetic flight vehicle based on experimentally measured wing kinematics. 3D optical motion capture, a standard method for obtaining wing kinematic measurements, discretely 6 samples the smooth surface of a bat wing during flight. If the constellation of tracked 3D points is too sparse, a loss of critical information occurs. Here, we have explored the relationship between the density of wing surface points and several aerodynamic metrics, specifically, wing surface area 1748-3190 variation, aerodynamic loads, and power expenditure. Loads and power were calculated using an incompressible Navier–Stokes solver. Of the metrics examined, aerodynamic power was found to be most sensitive to the spatial fidelity of the wing—the normalized root mean squared difference 1 (NRMSD) between the 10- and 238-point cases was 35%. Load calculations varied slightly less with a peak NRMSD of 24% between the highest and lowest fidelity cases. Lastly, the wing surface area was 16 least sensitive to the spatial fidelity of the wing kinematics, with a maximum NRMSD surface area of 8%. Close similarity in aerodynamic behavior was observed when using either a 120- and 238-point surface representation, establishing a bound to the sensitivity between wing shape and aerodynamics. The results from the 10- and 22-point configurations demonstrate that sparse representation of a wing surface can lead to a loss of information. The characterization of kinematic complexity of the wings both informs how many degrees of freedom are important to measure and also informs how many degrees of freedom are required to robotically reproduce the flapping flight. 1. Introduction Of the three taxa capable of sustaining active flight—insects, bats, and, birds—the former two exhibit As the interest in small-scale flight has risen in recent strongest dependence on unsteady aerodynamic effects. years, biological organisms have emerged as natural Researchers initially showed that the leading edge vortex models for engineered flight vehicles (Shyy et al 2010). (LEV) enhances lift by delaying stall during insect flight Despite progress, a comprehensive understanding of (Ellington et al 1996), and subsequently discovered animal flight biomechanics remains incomplete (Chin several additional unsteady mechanisms (Dickinson and Lentink 2016). Early research into animal flight 1999). While a substantial amount of progress has been applied quasi-steady models borrowed from fixed- made towards the understanding of insect flight (Sane wing theory. However, it was quickly recognized that 2003, Wang 2005, Shyy et al 2010), bat flight has received unsteady aerodynamic effects caused by wing flapping comparatively less attention (Chin and Lentink 2016). must be considered to obtain a clear understanding of The complex morphology of a bat wing makes 24 animal flight (Maxworthy 1981, Ellington 1984). the investigation of bat flight technically challenging. September © 2019 IOP Publishing Ltd 2019 Bioinspir. Biomim. 14 (2019) 066011 P Windes et al In contrast to insect wings or bird wings, bat wings eton pose. In 2018, Windes et al conducted a numer- have a flexible actively controlled membrane, and an ical simulation of a Pratt’s roundleaf bat (Hipposideros articulated skeletal structure (Swatz et al 2005). Dur- pratti) in straight flight using kinematic data consist- ing early research, it was hypothesized that bats were ing of 108 discrete marker points distributed over par ticularly adept at rapid maneuvers and load-car- both wings, including approximately 40 marker points rying due to their high wing surface area to weight measuring the deformation of the wing membrane ratio, articulated skeletal structure, and flexible wing (Windes et al 2018). membrane (Hartman 1963, Aldridge 1987). In 1987, Detailed wing kinematic data, representing time- Norberg and Rayner conducted an extensive study varying boundary conditions in the surrounding qualitatively correlating aspects of bat morphology air flow, is a strict prerequisite to conducting physi- to various flight capabilities—namely, speed, maneu- cally relevant aerodynamic simulations. However, the verability, agility, efficiency, and load carrying ability stretching, twisting, and folding of bat wings during (Norberg and Rayner 1987). flight make the measurement of kinematics challeng- Since Norberg and Rayner’s study, researchers ing (Bergou et al 2011). In 1975, Norberg used Kodak have been attempting to quantitatively link the unique 16 mm high-speed film to record lateral, ventral, and physiology of bats to their impressive flight capabili- frontal 2D projections of a Long-eared bat (Plecotus ties with detailed aerodynamic explanations, yet much auritus) in straight flight. The projected location of 13 progress remains. In 2008, Muijres et al, used particle discrete points—3rd to 5th digits, wrist, elbow, ankle, image velocimetry (PIV) to show that the LEV was pre- and tail—were measured every 1/100 s during the sent and significant in lift production during bat flight 2.35 m s−1 flight. Several years later, a significant break- (Muijres et al 2008). Power expenditure and flight effi- through was made by Rayner and Aldridge in 1985, ciency of bats have been explored by several researchers when they incorporated multiple cameras to directly (Hedenström et al 2009, Muijres et al 2011, 2012, von measure wing motion during flight using 3D recon- Busse et al 2013, 2014). Additionally, load-carrying struction (Rayner and Aldridge 1985). In their paper, ability (MacAyeal et al 2011) and maneuvering flight they provided a framework for capturing wing kine- (Rayner and Aldridge 1985, Aldridge 1987, Iriarte-Diaz matic data including filming, 3D reconstruction, data and Swartz 2008, Henningsson et al 2018) have received smoothing, and interpolating data voids. However, some attention, but an understanding of the relevant during their study only a single 3D point representing aerodynamic mechanisms remains to be discovered. the trajectory of the body of the bat was tracked. A particularly underdeveloped aspect of bat flight The advent of low cost, high quality digital cameras research, is the area of aerodynamic modeling (Heden- and more powerful computing has aided researchers strom and Johansson 2015). Numerical simulations in obtaining better 3D kinematic data from bat flight. are a powerful tool to analyze complex unsteady fluid A variety of techniques for measuring 2D and 3D kin- flows, such as the flow of air around animal wings. In ematics of biological systems were outlined by Hedrick 1996, Smith et al applied the unsteady panel method (2008), and have been applied to bat flight by several to the study of hawkmoth flight (Smith et al 1996), researchers (Tian et al 2006, Hubel et al 2010, 2012, Wolf and shortly after, Liu et al used a full Navier–Stokes et al 2010, Muijres et al 2011, 2012, Riskin et al 2012, Von solver to further investigate the same (Liu et al 1998). Busse et al 2012, von Busse et al 2014). The temporal Since then, the use of computational tools has greatly resolution (measurements per second) and spatial reso- enhanced our understanding of insect flight aerody- lution (points measured per frame) of wing kinematic namics by providing researchers exceptional spatial measurements have varied widely from just a single and temporal resolution of air pressure and velocity point tracking the center of mass, to over a hundred fields in proximity to the wing. Numerous compu- points measuring detailed membrane deformation. tational studies have been conducted on the flight of Ideally, the spatial resolution captured during the cicadas (Wan et al 2014), dragonflies (Vargas et al 2008, kinematic measurements should correspond to the Koehler et al 2012, Li and Dong 2017), hummingbirds degrees of freedom of the wing. Risken et al, meas- (Song et al 2014, Ren et al 2016), and

Load more