Hyun Research Program: GEMUS Flapping Wing Vehicle

Deliverables

Premise

Under the advising of Dr. Nak-seung Patrick Hyun and alongside Jeffery Zhou, Gilbert Chang, and Logan Dihel, a gliding-enabled, mode-switching, unmanned sensing (GEMUS) flapping wing vehicle is being developed as a control system development platform to better understand how birds with high aspect ratio wings are able to conserve energy when faced via dynamic soaring. This is of particular interest because flapping wing vehicles have not previously been experimentally proven to be capable of dynamic soaring via the Rayleigh cycle. A large-scale robot with a high aspect ratio would be capable of high endurance during flight thanks to its ability to both glide and flap effectively without requiring high flapping frequencies, which are characteristics seen by seagulls and albatrosses. The end goal is to ultimately produce a platform that can "harvest" energy from wind disturbances to some extent to create the most energy-efficient flapping wing vehicle. The preliminary design has been partially validated through force transducer tests, with wind tunnel tests upcoming for further validation after further mechanical optimization is achieved.

Development

The current flapping wing vehicle prototype, denoted as version 2, has a wingspan of 1.215 m and a chord length of 0.3094 m. The semi-elliptical wing was fabricated by using a printed template from a wide-format printer to guide a rotary fabric cutter through ripstop nylon fabric. The template, designed by me, is a CAD drawing with hidden edges visible to mark accurate carbon fiber spar placements for assembly of the wing. The paper template, along with 3D printed spacers for the spars and nylon repair patches as adhesive strips allow for a wing to be fabricated and assembled in around 2 hours. In the future, a new template will be utilized that has a span approaching 1.5 m instead. An XFLR5 panel method analysis drove us to select the AS6098 airfoil to serve as the rib cross-section near the wing root, as it appeared to perform the best at high angles of attack which will allow for the robot to perform aerobatic maneuvers. The rest of the wing is effectively a flat plate, in which the lack of a dedicated rib allows for immense amounts of twisting which bodes very well for thrust production during flapping.

Version 2's wings are directly driven by two KST X15-1809 servos, which were specified through an Ansys Fluent k-ω transient solver that allowed us to estimate the aerodynamic torque exerted on the wing and thus determine how much torque was necessary to flap the wings. All structural components are currently additively manufactured via fused deposition modeling with PLA filament, but waterjet cutting will eventually be utilized to fabricate carbon fiber components in a future iteration. The robot is controlled using the CrazyFlie system, while force transducer data is collected from an ATI Mini40-E.

Outcomes

GEMUS is still early in the development phase, but has proven to be effective at wing flapping. In the most recent force transducer test for version 0, the robot was able to produce 515 grams of thrust and 180 grams of lift. With the robot weighing around 250 grams, the lift-to-weight ratio is less than 1. However, at non-zero angles of attack, GEMUS would, in theory, easily be capable of producing enough force to stay airborne because of the additional vertical contribution from the thrust vector.

After optimizing the wing flapping mechanism to maximize force production, the robot will undergo subsonic wind tunnel tests to better understand flapping performance in non-zero freestream velocity.

My involvement with the project ended as the scope shifted more towards mechatronic design and away from optimization of aerodynamic behavior, as the current wing structure is believed to be nearly optimal for the proposed operating conditions.